An organization's financial resources (e.g., cash)
available for investment in projects
or other uses.
The portion of the capital budgeting process that
deals with the allocation of the organization's available financial
resources across business units, programs, and projects. The goal is to allocate resources in
such a way as to maximize the value
derived from the expenditure of those resources. Capital allocation is
difficult in part because of the many alternative ways that specified
budget can be allocated across different entities and because of the
difficulty of determining how much benefit would be produced under each
option. Optimizing the allocation of capital is a primary goal of
project portfolio management.
The process used by an organization to select and plan
the expenditures that it will undertake over some defined, upcoming time
period, for example, the next quarter, next year, or next five years. Such
expenditures might include investments in property, plants, equipment,
R&D, advertising, and so forth. For project based organizations,
a key component of capital budgeting is deciding which specific projects to conduct and how much to spend on
each. Since some expenditures may be quite large, capital budgeting may
include determining how to finance those expenditures. These are important
decisions, as they impact risk and the
future success of the organization.
Traditionally, to support the selection of projects for
the capital budget, organizations have mainly relied on techniques for
estimating the incremental financial benefits available from each proposed
project, expressed, for example, as a net
present value (NPV), internal rate of
return (IRR), profitability
index (PI), and payback
period. Project portfolio management
(PPM) can improve on traditional project selection techniques by
accounting for other types of project benefits in addition to financial benefits and
by identifying project portfolios that collectively generate the most value
for their costs. Organizations that are implementing PPM typically do so
such that portfolio analyses are completed and the recommended project
portfolio is identified for decision makers as the capital budget is being
defined (see this paper for a description of
how PPM is implemented in support of capital budgeting).
A measure of an individual's subjective preference or
satisfaction expressed numerically on a cardinal scale. This means that the utility numbers
assigned (e.g., 1, 2, 3, etc.) are measured with respect to a benchmark
scale, a scale such that both the numbers and differences between the
numbers are meaningful measures.
A cardinal utility measurement is
analogous to a measurement of temperature using a thermometer—both are expressed on cardinal scales. A
thermometer measures temperature on a benchmark scale because the numbers
for temperature obtained from the thermometer tell you both whether today
is hotter than yesterday and by how much. It doesn't matter whether the
thermometer measures in Fahrenheit or centigrade, because the two are
linearly related (degrees Fahrenheit equals degrees centigrade times 1.8
plus 32). Likewise, a cardinal utility measure of preference will tell you
both whether you prefer one thing to another and by how much. Like the
number obtained using a thermometer, the utility number assigned is unique
up to a linear transformation. Contrast cardinal utility with ordinal utility.
cardinal utility function
A utility function that expresses
utility on a cardinal scale.
cardinal value function
A value function that expresses utility on a
cardinal scale. A cardinal value function is also called a measurable
As further explanation, a value function is a special case of a utility function
applicable only to conditions of certainty (i.e., when there is no uncertainty regarding the levels of the
utility function's attributes). Thus, a cardinal value function is the special case of a
cardinal utility function
applicable when there is no uncertainty over the levels of the attributes.
Also called certainty equivalent or risk-adjusted value, a measure of the value of an alternative or outcome
that takes uncertainty (risk) into account. The certain equivalent of an
alternative is the amount that the decision maker would be indifferent
between (1) having that monetary amount for certain or (2) having the
alternative with its uncertain outcome. The certain equivalent for a
project is typically less than its
expected value and depends in part on the
decision maker's willingness to accept risk. For example, a risk averse
decision maker might have a certain equivalent of $500,000 for a project
with equal chances of yielding $0 and $2,000,000, even though the expected
value for this project would be $1,000,000.
theory, which provides a means for encoding a decision maker's
preferences in a mathematical utility
function, provides a means for calculating the certain equivalent. A
utility function U provides a number that quantifies how satisfied the
decision maker would be, depending on the outcome. These utility functions
can be indexed by risk
tolerance, which quantifies the decision maker's willingness to accept
risk. The greater the decision maker's risk tolerance, the closer the
certain equivalent of a gamble will be to its expected value.
Among the useful properties of the certain equivalent is
the fact that the certain equivalent of the sum of lotteries is the sum of the
certain equivalents of each lottery, provided that the uncertainties are independent
(not correlated). This means that, just as the
of uncertain independent projects may be added to obtain the expected
value of the project portfolio,
if the decision maker is risk averse, the certain equivalent of each project may be added to obtain
the certain equivalent of the project portfolio.
characteristic objects method (COMET)
A multi-criteria analysis
(MCA) decision-making method based on paired comparisons and concepts of
fuzzy logic. A criticism of many MCA
methods using pairwise comparisons is rank reversal—adding a new
alternative to the original set can cause the ranking of the other
alternatives to change. COMET avoids this problem by applying pairwise
comparison not to the original set of alternatives but to a larger set of
constructed alternatives called characteristic objects which are
independent of the actual decision alternatives. Pairwise comparisons with
respect to each of the criteria are used to obtain a relative preference
measure for each of the characteristic objects. Preferences for each of the
original decision alternatives are then obtained based on a mathematical
concept of its distance from the nearest characteristic objects and their
A test that may be used to verify that a performance measure used to compare
alternatives is an observable; that is, capable of being observed and
measured. Being observable is a desirable characteristic for the measures
used in a decision model.
Suppose you are a decision maker in a company considering
a change to one of your products. One objective might be satisfying your
customers. Is "customer satisfaction" a useful performance measure?
To conduct the test, imagine the existence of a
clairvoyant (someone who can foresee the future). Assume you decide to make the
specified change to you product. Ask the clairvoyant to tell you the level
of customer satisfaction that will result. Could the clairvoyant answer
your question? The answer is no, because the clairvoyant would need to ask
you to define customer satisfaction and how you want to measure it. If, on
the other hand, your performance measure was "the number of customer
complaints received" or your company's ranking in the next customer
satisfaction survey, the clairvoyant could tell you that. Performance
measures that fail the clairvoyant test are not observables. Performance
measures that aren't observables are ambiguous or vague so should not be
used in decision models.
As in client-server, see Server.
A communications network. The word most often refers to
the Internet, and more precisely to some data center with servers that is
connected to the Internet.
An overused buzzword that can mean different things.
Originally, the term referred to a technology that mimics supercomputing
capability (high-speed mathematical calculations) by simultaneously
applying the computing resources available from numerous servers available
on a network. The concept is to divide a computational task into pieces and
use specialized connections to allow groups of servers (typically those
involving low-cost PC technology) to simultaneously conduct the data
Today, the term most often refers to a method for
delivering information technology services
over the internet through web-based
tools, rather than through personal computers or direct connections to
local severs. Cloud computing relies
on sharing computing resources. Rather than having to build and maintain
computing infrastructure in-house, organizations can purchase the desired
computational resources as a utility, just as they purchase water and
Some project portfolio
management tool providers, and other software vendors, are using the
term in a more mundane way as a synonym for software as a service (SaaS)—their
applications are available via the Internet cloud provided from their
Web servers and accessed on demand via the user's Web browser.
comfort zone biases
A category of related cognitive biases with the common
characteristic that their effect is to promote behavior that is comfortable
rather than reasoned. Examples of comfort zone bias include status quo bias, supporting evidence bias, and
sunk cost bias.
The process and steps for introducing a new product into
the marketplace. Decisions around commercialization are often critical to
market success, including when, where, and how to launch the product.
Commercialization is often the most expensive component associated with
projects designed to create new consumer products, as it typically includes
the cost of advertising, sales promotion, and other costly marketing
A multicriteria analysis, decision-making method with the characteristic
that poor performance
relative to one or more criteria can be compensated for by good performance relative to
other criteria. Compensatory methods typically involve four steps:
- Specifying criteria for
- Assigning weights to indicate the
importance of the criteria
- Judging the performance of each alternative with respect to each
- Applying an aggregation
equation or mathematical algorithm to combine the weights and
performance judgments to compute a "desirability" measure (number) for
ranking the alternatives
Examples of compensatory methods include multi-attribute utility analysis (MUA),
the analytic hierarchy process (AHP), the evidential reasoning
approach (ER), and most scoring models.
Project portfolio management tools with the capability to prioritize project almost
always use a compensatory method.
Also called compromise method, a sub-category of
multi-criteria analysis (MCA) decision-making methods composed of methods that rank
the alternatives to a decision based on calculating how "close" the alternative is to an ideal solution and
how "far" it is from the nadir solution. The ideal alternative is defined as a hypothetical alternative that
performs at the best possible level on every criterion. The nadir alternative is defined as a hypothetical alternative that performs at the
worst possible level on every criterion. "Distance" is defined using the natural extension to the concept of distance in 3-dimensional
space to the multi-dimensional space defined by the number of criteria. VIKOR
and the Technique for Order Preference by Similarity to Ideal Solution (TOPSIS) are examples
of compromising methods.
Methods in this category are termed "compromising" based on their interpretation when applied to high-conflict,
group decision making situations.
Presumably, stakeholders can agree on the criteria for evaluating alternatives. Thus, everyone should agree on the ideal and nadir
solutions. An alternative that is as close as possible to the ideal and as far as possible from the nadir should be one
that achieves a high level of agreement. The group as a whole should be pleased that the selected solution is close to the ideal. Those
opposing the choice should presumably feel minimal regret based the choice being so far from the worst possible choice.
Also called concordance analysis a sub-category of multicriteria analysis (MCA) methods based on
identifying and comparing criteria that support a particular alternative (concordance) alongside the criteria that oppose the
alternative (discordance). (Concordance means agreement, concord, harmony.) Concordance methods are a type of outranking method.
The various ELECTRE methods and PROMETHEE are the most used and well-known of the concordance methods.
Like other outranking methods, with concordance methods each pair of alternatives is compared to determine
whether one alternative outranks another (performs as well or better on every criterion). Once all alternatives have been
compared, the results are combined to obtain a partial (or complete) ranking order.
Concordance methods are non-compensatory in that there is no need to assess strengths of preference or
tradeoffs among the criteria; all that is required are the ordinal results from the pairwise comparisons. The limitation is that
the resulting ranking is ordinal rather than cardinal, meaning that there is no ability to divide a utility number by cost to
obtain a ranking that maximizes portfolio value..
A range of possible values defined for an uncertain
quantify such that there is some specified probability (e.g., 90%) that the
true or actual value of the quantity lies within the range. The term has a
somewhat different meaning in statistics, where it describes an interval or
range for some statistic computed from a sample of observations from some
population (e.g., the mean value) such
that the "true" population statistic can be expected to be located within
that interval with specified level of certainty (e.g., 90%).
A method often used in the field of market research to
quantify how people value the various features or attributes that make up a
product or service. Subjects are asked to state their preferences for
choices presented as paired comparisons of
product choices defined by two or more attributes, for example, a low cost
computer with 2 GHz processor, 512 MB RAM, and a 15 inch monitor versus a
higher cost computer with 4 GHz processor, 2 GB RAM and a 21 inch monitor.
In some versions, the subjects indicate their strength of preference, for
example, on a 1 to 10 scale. The approach requires answers to be provided
for many such questions, however, since most people find the questions
easy, conjoint studies can often be conducted by mail survey or over the
The data obtained from a conjoint analysis may be used to
derive a utility function for
quantifying customer preferences for products with various combinations of
attributes. A specific mathematical form for the utility function is
assumed (e.g., additive) and the data from the conjoint analysis is used to
set the parameters of the function so as to best fit the data. The utility
function may then be used to predict future customer choices. The technique
may be used to quantify preferences for items other than products, and some
project prioritization tools
use conjoint analysis as a method for ranking projects.
A type of a scoring scale designed to measure the connotative meaning of objects, events, or concepts.
The connotations are used to derive a subject's attitude towards the given object, event or concept scale.
On this website, the term typically refers to a decision to conduct a
project. Some projects, for example, may have positive, negative, or neutral connotations.
An outcome of a decision. On this website, the term typically refers to a decision to conduct a
Consequences are outcomes that matter to the decision maker; that is, an outcome that the
decision maker cares about. Consequences are typically
quantified via performance measures that indicate the degree to which
the decision maker's objectives are achieved.
A model for simulating or otherwise relating results or
outcomes that people care about to specified actions and conditions.
Project portfolio management tools that
prioritize projects based on the likely consequences of those projects
contain consequence models. Consequence modeling is the process of
constructing a consequence model and typically involves developing a
description of cause-effect relationships or linkages. Various techniques
are available for creating consequence models, and many such models exist
and have become well-established within various disciplines and application
In risk analysis, a policy of assuming pessimistic
outcomes for uncertainties when quantifying risk. Conservatism reflects a
preference for erring on the side of overstating as opposed to understating
risk in situations where uncertainties complicate the quantification of
risk. For example, if a probability distribution for
an uncertain undesired outcome is known, conservatism might involve selecting an outcome
estimate at, say, the 95th percentile—meaning there is a 95% chance
that the actual outcome will be overestimated (the outcome will be more desirable than assumed) and only a 5% chance that it is
underestimated (less desirable than assumed). By substituting conservative estimates for probability
distributions, conservatism allows for a simpler, deterministic analysis of
risk. Conservatism is criticized, however, for injecting a bias into risk
analysis. Also, studies show that conservative assumptions tend to compound
with the result that estimated risks are far larger than an objective
analysis would report.
In the mathematical sense, a constraint is a condition of an decision problem that
the solution must satisfy. There are several types of constraints—primarily equality constraints, inequality constraints,
and integer constraints. The set of candidate solutions that satisfy all constraints is called the feasible set
Constraints may be dictated by the environment, physical processes, economics, cultural,
legal and/or resources aspects which must be satisfied in order to
obtain an acceptable solution. As one example, conducting all projects required by law is a common constraint
imposed on the choice of a project portfolio.
A class of mathematical solution techniques aimed at
optimizing an objective
function with respect to variables some of which are subject to
constraints. The techniques for constrained optimization include linear programming, non-linear programming,
integer programming, dynamic programming, and
Also called a constructed measure, a measure
constructed in terms of other measures. Examples include the Dow Jones Industrial Average, the Consumer Price Index, and the
Michelin 3-star measure for rating restaurants. Constructed measures are typically defined when neither
natural measures nor proxy measures
cannot be found. Sometimes, the term is used to refer to a measure based on a constructed
See attribute and
The difference between the total amount that consumers would be willing to pay
and the total amount that they actually do pay (i.e. the market price). In other words, consumer surplus is the value consumers receive over
and above what they actually have to pay for goods and services. Consumer surplus is a type
of value obtained by customers when organizations provide products or services that are purchased by customers.
Resources set aside for expenses associated with
unexpected but potential project or
other business setbacks. For example, a contingency fund might be
established in the context of project
portfolio management (PPM) to provide additional funding for projects
that might go over budget (although the PPM process would include
reevaluating such projects to see if the additional funds are warranted and
what changes should be made to the project plan in light of a budget
A method, generally associated with cost benefit analysis, for assigning monetary
values to consequences based on asking individuals about their willingness
to pay to obtain desired outcomes or to reduce or willingly accept adverse
outcomes. Contingent valuation is often used to value environmental
consequences such as increased levels of pollution or noise.
continuous risk, continuous
A risk or uncertainty
with that will produce an outcome within a range of possibilities.
Continuous uncertainties are characterized by continuous probability distributions
(e.g., probability density functions) that indicate the relative
probability of the possibilities. For comparison, see discrete risk.
A measurement scale
that allows assigning an infinite number of possible number values, for
example, any number between zero and one.
A statistical relationship between two or more variables
such that systematic changes in the value of one variable tend to be
accompanied by systematic changes in the other. For example, as illustrated
below, the height of parents and mature children can be shown to be
correlated based on plotting on a graph parent heights versus child
heights. If there was no correlation, the plot would appear as dots spread
across a roughly circular shape. Instead, the plot appears as an oval shape
aligned along approximately a 45 degree line from the origin of the graph.
The shape demonstrates that taller parents tend to have taller children,
but the relationship is not perfect, indicating that there are other
A sample correlation plot (parent vs. children height)
Correlation may be quantified by calculating a
coefficient of correlation, denoted ρ, which is a number
between -1 and 1 that indicates the strength and direction of a linear
relationship between two variables. Regression analysis may be used to
compute correlation coefficients.
cost benefit analysis
Also called benefit cost analysis, a decision
making theory and collection of related analytic techniques for evaluating
decisions based on comparing benefits and costs. While many decision-aiding
tools claim to do this, CBA is distinguished by its foundation in theory
and by its rigorous process for computing costs and benefits. Like decision analysis (DA), CBA is
essentially a "megatool;" a coherent set of concepts and techniques that
may be used to identify, estimate, and place monetary values on the impacts
of proposed actions.
A distinct characteristic of CBA, compared to many other
decision aiding approaches, is that it does not quantify the costs and
benefits of actions from the specific perspective of responsible decision
makers. Instead, CBA measures costs and benefits from the perspective of
society at large. Thus, CBA is most applicable for government decision
making. CBA requires identifying all parties affected by a proposed action
and estimating the monetary value of the effects the action would have on
their welfare. The US government began using CBA in the 1930s, and CBA
continues to be the mostly widely used approach for formally evaluating and
prioritizing major projects undertaken by government agencies.
With CBA, costs and benefits are measured relative to a
"do-nothing," status quo option. According to CBA, an action should be
considered only if its net benefit (benefit minus cost) is greater than
zero. According to CBA, all alternatives with positive net benefit should
be considered, and the best alternative is the one that leads to the
greatest benefit gain (i.e., the alternative with the largest net
To determine the monetary value of the impacts of
proposed projects, CBA relies on the
concept of market prices. In theory, a free market generates prices by
balancing aggregate demand with aggregate supply. Each individual adjusts
his or her purchases until the value of the last item purchased is just
worth what it costs. Thus, the prices that result indicate the marginal
benefit realized from each individual's consumption of each good.
Following this logic, CBA attempts to use market prices
to value project impacts. For example, suppose a project is proposed to
clean up a hazardous waste site. CBA might use real estate prices to
estimate the value of the cleanup effort. The market values of similar
properties close to and far from the waste site would be compared to
determine the value loss suffered by those living near to the site.
Removing the site would, presumably, eliminate the property value
differences and create this much added value for home owners. This monetary
value could be compared to the costs of the project to determine whether
the project should be conducted. To value project impacts for which no
market exists, CBA uses procedures that indirectly reference market prices.
For example, values for CBA are often obtained based on surveys or
interviews with people to estimate their willingness to pay to obtain or
avoid the effects in question.
Although CBA is widely used and based on compelling
theory for decision making, the approach is considered by many to be
controversial. CBA ignores the way in which the costs and benefits of
proposed actions are distributed, a consideration that is often important
to decision makers. Also, numerous applications of CBA have demonstrated
that the approach frequently concludes that the benefits of proposed
government interventions do not justify the costs. This result may be due
the fact that some project benefits simply cannot be addressed through
reference to prices that exist in the marketplace.
Since CBA does not normally rely on methods for assessing
and incorporating expert judgments, it can fail to account for project
outcomes that do not have immediate, tangible, economic implications.
Likewise, CBA may fail to address important risks in situations lacking
data for quantifying uncertainties. CBA provides little opportunity for
stakeholders to contribute to the evaluation process, except perhaps, in
framing the problem (e.g., identifying alternatives). On the other hand,
CBA avoids the necessity of decision makers providing subjective value
judgments. It therefore appeals to some because it appears to be a more
value-free guide to decision making. In truth, though, CBA embodies strong
cost effectiveness analysis (CEA)
A form of economic analysis that compares the costs of
alternatives for achieving the same or similar outcomes. CEA can be viewed
as a version of cost-benefit analysis
(CBA) that does not include monetizing the estimated benefits of alternatives, and, as such, it is
often used in situations where it would be difficult or uncomfortable to
put a dollar value on outcomes.
For example, "number of lives saved" might be an outcome
(or performance measure)
for investments in health care. In order to compare or prioritize health
care investments, CBA includes techniques for expressing outcomes in
equivalent monetary units so that the values can be combined and compared
with costs. (Other decision models also express the value of various
outcomes in common units, though the unit need not be dollars.) CEA might
be preferred over CBA if assigning an explicit monetary value to saving a
life is viewed as too controversial.
Though CEA may be less controversial than CBA, the
technique is not so helpful if the alternatives under consideration have
different costs and would produce different outcomes. If, for example, more
lives would be saved under a more expensive alternative, then decision
makers must judge whether the additional lives saved from the more
expensive alternative would be worth the additional costs. In the absence
of a model that allows assigning explicit values, those tradeoffs are made
implicitly, so the logic for choice is less transparent. For this reason,
CEA is most useful when the alternatives under consideration have different
costs, but would produce the same outcomes.
cost of capital
The cost a company effectively pays in order to obtain
cash (from debt and equity sources) to finance its operations. It is the
return that is required on company investments to compensate the investor,
taking into account the risk involved. The cost of capital is generally
calculated on a weighted average basis (see WACC).
Any quality, property, characteristic, or rule
established to guide decision making. The term has somewhat different meaning in different contexts, as attributes, objectives and goals may all be referred to as criteria. Reflecting this, the term a multicriteria
decision problem means either a multi-attribute or a multiobjective decision problem (or both).
The term multicriterion decision making (MCDM) is, therefore, used to
indicate the general field of study which includes decision making in the presence
of two or more conflicting objectives and/or decision analysis processes involving two or more attributes.
of multiple criteria for decision making is the subject of multi-criteria analysis.
The subset of activities within a project whose duration determines the duration
of the project. If an activity along the critical path is delayed, then the
project will be delayed.
A probability distribution
expressing the probability that a random variable X will take a
value less than or equal to x. The figures below provide examples of cumulative probability distributions. The first is
as sample cumulative probability distribution for a discrete uncertainty , the second is for a continuous uncertainty.
The internal rate of return (IRR) for a project below which the project is considered unacceptable. It is often taken to be the project
opportunity cost or the organization's cost of capital. The cut-off rate would be the
minimum acceptable irr for a project or the discount
rate used to calculate the project's net present value.